Application of artificial intelligence techniques to energy management in hybrid system

Abstract

Meeting the growing global energy demand requires innovative and sustainable solutions, with renewable energy resources (RERs) offerin g apromisin gpathway .Among these, photovoltaic (PV) systems stand out as accessible and scalable options, especially when integrated at the distribution level in smart grids. However, the inherent variability and unpredictability of solar energy introduce significat challenges for real time energy management and grid stability. Accurate forecasting of power production and consumption thus becomes a critical task in ensuring optimal energy flow and efficient grid operation. This study presents a case-based investigation of a grid-connected PV system enhanced with artificial intelligence(AI) techniques for advanced energy management. Specifically ,it explore sand compares the performance of an Adaptive Network based Fuzzy Inference System (ANFIS), a Convolutional Neural Network (CNN), and a hybrid CNN–ANFIS model in predicting short-term power generation and consumption. All models were developed, trained, and evaluated using MATLAB, based on real world weather and energy data collected over multiple seasons. The evaluation metrics, particularly root mean square error (RMSE), demonstrate that the hybrid model combines the learning capabilities of CNNs with the reasoning strength of ANFIS, offerin gimproved forecasting accuracy. The integration of such intelligent predictive systems supports smarter decision-making and can significantly enhance the reliabiliy and cost efficiency of distributed renewable energy integration in modern power systems.